Skip to content

Lab 24 — Managing a Distributed Team

At a glance

Learning objectives

By the end of this lab you should be able to:

  1. Identify the coordination costs that distribution adds, and express them in hours.
  2. Design a working agreement that substitutes for the context a co-located team gets for free.
  3. Make site boundaries and interface ownership explicit rather than implicit.
  4. Correct for visibility asymmetry deliberately.

Scenario

This lab continues the Admission Management System from Labs 08, 09, and 10. The university has insisted the work be split across two sites to build capability in the regional campus.

Item Value
Site A Main campus: 5 people — the project manager, the business analyst (60%), 2 developers, the tester (50%)
Site B Regional campus: 4 people — 1 team lead, 2 developers, a database administrator (50%)
Time difference Site B is 4 hours behind Site A
Overlap 3 hours of shared working time daily
Duration 40 working days (from Lab 10's network)
Work split Site A owns the applicant form and eligibility rules; Site B owns the admin console and the student-record integration

Two problems have already surfaced in the first three weeks:

  1. The API contract between the applicant form (Site A) and the integration (Site B) has been adjusted locally by both sides for convenience. The two versions now differ in how they represent the applicant's programme choice. Neither site considers its version wrong.
  2. A design disagreement about how the eligibility rules are surfaced to applicants was discussed in the three-hour overlap and, by Site A's account, concluded; by Site B's account, still open. Nothing was written down.

Team observations recorded this week:

  • Daily written updates were agreed at the start of the project. Four of the nine people send them.
  • Two people have never spoken to anyone at the other site.
  • Site B's work produces no visible interface until integration, so it is barely mentioned in the fortnightly progress review; Site A's form screens are shown at every review.
  • The security review (activity H) is owned by a third party at the main campus and does not know that Site B adjusted the interface.

Your role

You are the project manager, based at Site A. The steering committee has asked for a revised working arrangement, and it wants to know what the arrangement costs. You must produce it.

Project information and assumptions

The scenario is part of the example-project library — read the full description there if you want the wider context.

  • Use 40 hours per week as one full-time equivalent (FTE).
  • All nine people are counted as participants, with the analyst at 0.6 and the tester and DBA at 0.5 where their availability matters.
  • The three-hour overlap is the only synchronous window for the whole team; it is scarce and must be allocated deliberately.
  • Assume the two-site split is a decision that cannot be reversed in this project. Your job is to make it work, and to say what it costs.
  • Assume the design disagreement is genuine: both accounts are honestly held.

Tasks

Task 1 — Map the coordination surfaces

List every point at which work must pass between the two sites. For each, record:

Surface What crosses Who owns it Current practice Failure mode observed

At least one surface must be a contract or interface, one a decision, one a handoff of work, and one a dependency on a third party.

Task 2 — Diagnose the two problems

For each of the two problems described:

  1. State the underlying cause, in terms of the coordination surfaces from Task 1.
  2. State what a co-located team would have that this team does not, which would have prevented it.
  3. State the concrete artefact that would have prevented it, and describe its content in three or four lines.

Then answer: which of the two problems is more serious, and why? Consider both the cost of fixing it and the cost of not noticing it.

Task 3 — Write the working agreement

Write a working agreement for this team with at least eight clauses. Every clause must be checkable — someone should be able to say whether it is being followed.

Cover, at minimum:

  1. How decisions are recorded, by whom, and within what time.
  2. When a live session is justified, and when it is not.
  3. How the daily update works, and what happens when it is missed.
  4. Who owns the API contract, and what must happen before it changes.
  5. How a design disagreement is closed.
  6. What the overlap window is reserved for.
  7. How a new team member is onboarded at either site.
  8. What happens to unfinished work at the end of a working day across the time difference.

Then state which clause you would enforce first, and why enforcing that one changes the others.

Task 4 — Make ownership explicit

Complete an ownership map:

Interface or component Owner (role and site) Consumers Change process Escalation route

Include the API contract, the data model, the shared test environment, and the security review.

Then answer: what specifically went wrong because the API contract had no named owner, and what would ownership have changed in week 1 or 2?

Task 5 — Design the handoff artefact

Site B's integration work depends on the data model produced at Site A, and the handoff currently happens in conversation during the overlap.

  1. Design the handoff artefact: its name, its contents, its author, and where it lives.
  2. State what it must contain for a person who has never spoken to the author to act on it.
  3. State how it is versioned and what happens when it changes.
  4. Estimate the author's effort per handoff, in hours, and the number of handoffs in the 40-day plan.

Task 6 — Cost the coordination overhead

Using the figures below — which you should treat as assumptions for teaching — calculate the weekly overhead of the working arrangement and express it as a fraction of an FTE.

Practice Cost
Daily written update 15 minutes per person per day, all 9 people
Two coordination sessions per week 1 hour each, 6 attendees each
Weekly interface review 1 hour, 4 attendees, once per week
Handoff artefact authoring 1.5 hours per handoff, 4 handoffs in the 40-day plan

Complete:

  1. The weekly figure for each practice.
  2. The total per week.
  3. The total as a fraction of one FTE (40 hours per week).
  4. The total over the 40-day plan (8 weeks), in hours and in FTE-weeks.

Then answer:

  1. Is any of the four practices a candidate for removal? Justify with a number and a consequence.
  2. If the time difference were 8 hours with no overlap, what would happen to the synchronous practices, and what would you replace them with?
  3. Which practice is the most expensive, and does its cost surprise you? What does that say about where coordination effort actually goes?

Task 7 — Correct the visibility asymmetry

Site B's work is barely mentioned in progress reviews. Design the correction:

  1. State exactly what you will change in the fortnightly progress review.
  2. State who presents Site B's work, and why it should not be the project manager.
  3. Explain what under-recognition does to a remote site over a long project, without relying on motivation alone — include at least one consequence that affects delivery.

Task 8 — Update the security review

The security review does not know the interface changed.

  1. State the risk this creates, in cause–event–impact form.
  2. State the corrective action, with an owner and a date.
  3. State what would have prevented it, in one sentence, referencing Task 4.

Deliverable

  1. The coordination surface map.
  2. The diagnosis of both problems, with the artefact that would have prevented each.
  3. A working agreement with at least eight checkable clauses, plus the enforcement priority.
  4. The ownership map with the four required entries.
  5. The handoff artefact design with its effort estimate.
  6. The coordination overhead calculation and all three answers.
  7. The visibility correction and the security review update.

Hints and guidance

  • A written decision reaches everyone at once. In a team spanning four hours, anything decided verbally in the overlap is invisible to whoever could not attend.
  • Ownership is not a title; it is a change process. Naming an owner without saying what must happen before the interface changes leaves the problem exactly where it was.
  • Async practices cost time, and the cost is worth stating. 0.68 FTE sounds wasteful until you compare it with the cost of the API divergence you already have.
  • Under-recognition is a delivery problem, not only a morale problem. Work that is not seen is not resourced, and risks in it are not surfaced.
  • The overlap is the scarcest resource you have. Do not spend it on status; spend it on disagreement and negotiation.

Self-check criteria

Check How to verify it
Four coordination surfaces with all five fields No surface missing an owner
Both diagnoses name a written artefact Not "better communication"
Working agreement clauses are checkable Each can be judged followed or not followed
Ownership map includes the API contract With a named owner, consumers, and change process
Handoff artefact is actionable by a stranger Its contents are listed, not described in the abstract
Overhead arithmetic correct Daily updates 11.25 h/week; sessions 12 h/week; interface review 4 h/week; handoffs amortised
FTE fraction stated Total ÷ 40, with an interpretation
At least one practice questioned With a number and an accepted consequence
Visibility correction is specific It names a change to the review, not an intention

Reflection

  1. The daily written update is the single largest recurring cost. Would you remove it, and what would you put in its place?
  2. Ownership of the API contract was unassigned. Was that a planning failure or a communication failure? What process would have caught it?
  3. The design "agreement" was recorded nowhere. Would a written decision have settled the disagreement, or merely documented it? What else is needed?
  4. If the two sites were swapped — Site B holding the form and Site A holding the integration — would your working agreement change? What does the answer tell you about who holds power in a distributed arrangement?
Check your overhead

Weekly coordination cost

Practice Calculation Hours per week
Daily written update 9 people × 15 min × 5 days = 675 min 11.25
Two coordination sessions 6 attendees × 1 h × 2 12.00
Weekly interface review 4 attendees × 1 h × 1 4.00
Handoff authoring 1.5 h × 4 handoffs ÷ 8 weeks 0.75
Total 28.00

As a fraction of an FTE: 28 ÷ 40 = 0.70 FTE — that is, seven-tenths of one person's entire working week spent coordinating rather than producing.

Over the 40-day plan (8 weeks): 28 × 8 = 224 hours, or 5.6 FTE-weeks.

The dominant cost is the two coordination sessions (12 h/week), not the written updates. That is the opposite of most people's intuition, and it is why the sessions should be reserved for disagreement rather than used for status.

Full working is in the Advanced labs answer key.

Extension challenge

The university proposes moving to three sites: main campus, regional campus, and a partner institution in a time zone 7 hours behind, with no overlap at all. Recalculate the coordination overhead for an async-only arrangement, and state exactly which of your working agreement clauses survive and which must be replaced. Then answer: what is the maximum number of sites at which this working arrangement stops being viable with the team size given, and what would have to change for the answer to improve?